The Field, Everyone in the superintelligence field.

That number is who qualifies on this field under a strict bar. It is not a claim that only 7 groups on Earth work on the risk.

Prevent is reserved for standing main organizations whose primary public ask is to stop, pause, ban, prohibit, or put a moratorium on frontier or superintelligent AI. Most of the field writes rules, runs tests, funds research, or builds systems while the race continues. Those sit in Govern or Build. Endorse a pause once and you still do not land here if your brand is something else.

Most of the money and prestige still go to making the race safer, not to ending the race to ASI.

Prevent is the brand on the door: stop, ban, or prohibit ASI. Build is a frontier lab whose product is a general system on that path. Govern is the large middle of the fight: people writing rules, running tests, moving money, or publishing research while the labs keep going.

Filter at the top to search by name.

Their primary public ask is not a ban or prohibition on artificial superintelligence.

Encode campaigns for frontier safety bills. CAIS runs research and field-building. Policy shops that push stronger rules, or sometimes endorse a pause, still sit in Govern when stop-ASI is not the brand on the front door. Padding Prevent with every safety-adjacent shop would hide how small Prevent is.

Organizations only, and only ones whose primary work is the superintelligence fight. Not people, podcasts, one-off campaign pages, city chapters, accelerators that belong under a parent, or defunct groups. Not general AI ethics, multi-cause charity, or near-term product governance as the main job.

PauseAI national chapters roll up to PauseAI. An FLI accelerator does not get a separate tile from FLI. The field is a census of main actors by trajectory, not every satellite and sub-brand.

This directory is about artificial superintelligence and the race to it. The whole field uses that same object: stop ASI, manage the path to ASI, or build toward it.

Groups whose main job is bias, privacy, ordinary product rules, multi-cause charity, or general AI ethics belong in other maps. They can matter. They are not this fight. We keep them off so the thin Prevent box and the wide Govern band stay honest about who is actually in the ASI race.

We open the organization's own site and look for stop, pause, ban, prohibit, or moratorium language as the main ask. Third-party blurbs and category tags are not enough.

ControlAI is the textbook case. The name and some blurbs say “control,” but the homepage asks to prohibit the development of superintelligent AI, so it sits in Prevent. If a standing org clears that bar and is missing, tell us.